CN108206790A - A kind of selection of SDN joint routes and resource allocation methods based on network slice - Google Patents

Abstract

The invention relates to an SDN joint routing selection and resource allocation method based on network slicing, and belongs to the technical field of communication networks. The method is as follows: S1 SDN controller receives user stream transmission request; S2 controller determines candidate switch set and candidate link set according to user stream resource requirements; S3 models network slice and resource allocation identifier; S4: determines network slice and resource allocation Restricted conditions; S5 models the user flow end-to-end delay and network load; S6 aims to minimize the user flow end-to-end delay and network load, and determines the network slicing strategy for each user flow optimization; S7 judges whether each network slice There are shared switches and shared links; the S9 controller sends network slicing and resource allocation policies to corresponding switches, and each switch completes user flow routing and resource allocation according to the flow table. The invention aims at minimizing end-to-end time delay of user flow and network load, and realizes optimization of SDN joint route selection and resource allocation.

Description

A SDN joint routing selection and resource allocation method based on network slicing

technical field

The invention belongs to the technical field of communication networks, and relates to an SDN joint routing selection and resource allocation method based on network slicing.

Background technique

Software Defined Network (SDN) is an emerging software-based network architecture and technology, which can be applied to networks with large-scale and heterogeneous characteristics. The main features of SDN are the separation of the control plane and the data plane, logically centralized control of the network, and flexible and efficient network management and operation and maintenance through software programming. SDN separates the control plane from the data plane, which can realize the logical concentration of network control functions, which greatly simplifies the management and control of routing; the use of open application programming interface control can effectively shorten the service time of business flows and improve network resources. utilization rate.

In order to meet the transmission requirements of various types of business streams in the network, network slicing technology can be used, that is, on the basis of the same physical network, use technologies such as SDN and network virtualization to build multiple network slices on demand to meet different user needs. Effective support to improve user experience. Although SDN and network slicing can significantly improve network service support capabilities, the complex network architecture, heterogeneous integration of network forwarding equipment, and diverse service requirements in SDN networks have brought new problems and challenges to the routing mechanism and network resource allocation. challenge.

The literature [Alba Xifra Porxas, Shih-Chun Liny, and Min Luo.QoS-AwareVirtualization-Enabled Routing in Software-Defined Networks (ICC), 2015] proposes a multi-tenant management architecture in an SDN network, through tenant isolation, prioritization and Traffic distribution is used to meet the user's QoS requirements, but this literature does not consider the network load problem, which may cause some nodes or links to be overloaded. The literature [Su Xin, Zhang Li, Gong Jinjin, etc., a user service request selection method based on end-to-end network slicing, CN106210042A, 2016] proposes a user service request selection method for end-to-end network slicing. Users authenticate and authorize, and match appropriate network slices to respond to user requests, but this method does not consider the resource allocation ratio when multiple users share physical resources.

To sum up, in the SDN-based network environment, how to comprehensively consider the user's business needs, node, link capacity, and transmission characteristics, etc., to realize the user's joint routing selection and resource allocation scheme under the SDN architecture, thereby reducing network delay and balancing Network load becomes an urgent problem to be solved.

Contents of the invention

In view of this, the object of the present invention is to provide a method for SDN joint routing selection and resource allocation based on network slicing.

To achieve the above object, the present invention provides the following technical solutions:

A network slicing-based SDN joint routing selection and resource allocation method, comprising the following steps:

S1: a software defined network (Software Defined Network, SDN) controller receives a user stream transmission request;

S2: The controller determines a set of candidate switches and a set of candidate links according to user flow resource requirements;

S3: modeling network slicing and resource allocation identification;

S4: Determine network slicing and resource allocation constraints;

S5: Model user flow end-to-end delay and network load;

S6: Aiming at minimizing the user flow end-to-end delay and network load, determine the network slicing strategy for each user flow optimization;

S7: Determine whether there are shared switches and shared links in each network slice, if yes, perform step S8; otherwise, perform step S9;

S8: Determine whether the shared switch and the shared link meet the resource constraints of the switch and the link, and if so, execute step S9; otherwise, execute a resource allocation strategy for the shared switch and the shared link;

S9: The controller issues network slicing and resource allocation policies to corresponding switches, and each switch completes user flow routing and resource allocation according to the flow table.

Further, in step S1, the modeling of the user streaming request is Among them, S _k and D _k represent the source and destination switches of the k-th user flow respectively, and f _k represents the data flow of the k-th user flow, Indicates the minimum transmission rate requirement of the kth user flow, and Respectively represent the switch computing resource requirements and storage resource requirements of the k-th user flow.

Further, in step S2, the set of candidate switches is determined according to the switch resource requirements of the user flow, that is, N _i is used as a candidate switch for user flow k, only if the computing resources of the switch and storage resources Satisfy the resource requirements of user flow k, that is, and by Indicates the set of candidate switches for the kth user flow, namely

The set of candidate links is determined according to the link resource requirements of the user flow, that is, L _{i, j} is used as a candidate link of user flow k, only if the bandwidth resources B _i,j of the link meet the bandwidth resource requirements of user flow k , and the transmission rate R _i,j of this link meets the minimum transmission rate requirement of user flow k That is, f _k ≤ _{B i,j} and by Indicates the set of candidate links for the kth user flow, namely

Further, in step S3, the modeled network slice and resource allocation identifier are specifically: let x _i,j,k ∈{0,1} be the network slice identifier of user flow k, if x _i,j,k = 1, it means that the network slice allocated for user flow k includes switch N _i and link L _i,j , otherwise, x _i,j,k =0; respectively represent the proportion of computing resources and storage resources occupied by user flow k of N _i ; δ _i,j,k ∈[0,1] represents the proportion of transmission resources occupied by user flow k of L _i,j .

Further, in step S4, the network slicing and resource allocation constraints include switch flow conservation, switch, link capacity, and resource allocation constraints. Specifically, the switch flow conservation conditions are as follows:

Only one output path is selected at the routing source switch S _k of user flow k, that is,

The input flow and output flow at the relay switch N _j of the user flow k are equal, namely

Only one input path is selected at the destination switch D _k of user flow k, that is,

The switch capacity limiting condition includes the switch computing resource capacity and storage resource capacity limiting conditions, modeled as:

The computing resources occupied by N _i by the user flow cannot exceed the computing resources of the switch, that is, The storage resources occupied by N _i by user flows cannot exceed the storage resources of the switch, that is,

The link capacity limitation condition is that the link resource occupied by L _i,j cannot exceed the bandwidth resource amount of the link, that is,

The switch resource allocation constraints include switch computing resource and storage resource allocation constraints, modeled as:

The sum of the computing resource allocation ratios occupied by user streams N _i is not greater than 1, that is

The sum of the storage resource allocation ratios occupied by user streams N _i is not greater than 1, that is

The limiting condition for link resource allocation is that the sum of the bandwidth resource allocation ratios occupied by L _i,j for user flows is not greater than 1, that is,

Further, in step S5, the end-to-end delay of the system is modeled as the sum of link transmission delay and switch queuing delay, namely Among them, D _{i, j} represents link transmission delay, D _i represents switch queuing delay, and D _{i, j} and D _i are respectively modeled as follows:

Modeling link transmission delay

The data processing process of the modeling switch is an M|M|1 queuing system, and Among them, μ _i represents the business service rate of N _i , and λ _i represents the business arrival rate of N _i ;

The network load is modeled as the sum of link load and switch load, namely: Among them, U _{i, j} represents the link load, U _i represents the switch load, and U _{i, j} and U _i are modeled as follows:

Modeling link load

Modeling Switch Load in, Calculate the load for the switch, is the storage load of the switch, and α is the weight coefficient.

Further, in step S6, the network slicing strategy refers to determining the network slicing strategy for each user flow under the condition of satisfying network slicing and resource allocation constraints, that is, for each user flow k, set For the resource allocation strategy of each flow, with the goal of minimizing the end-to-end delay and network load of the user flow, optimize and determine the network slicing strategy, namely:

Further, in step S7, the judging whether each network slice has a shared switch and a shared link, that is, judging Whether it is established, and if it is established, there is a shared switch and a shared link.

Further, in step S8, the judging whether the shared switch and the shared link satisfy the resource constraints of the switch and the link are respectively:

Judging the computing resource constraints of the switch N _i , if the user flow occupies the computing resources of the switch exceeds its computing resources, that is Then, with the goal of minimizing the queuing delay of the switch and the calculation load of the switch, optimize and determine the calculation resource allocation strategy of the user flow at N _j

Judging the storage resource constraints of the switch N _i , if multiple user flows occupy the storage resources of the switch exceeding its storage resources, that is Then, with the goal of minimizing the storage load of the switch, optimize and determine the storage resource allocation strategy of the user flow at N _j

Judging the bandwidth resource constraints of link L _i,j , if multiple user flows occupy the bandwidth resource of the link exceeds the amount of link bandwidth resource, that is Then, with the goal of minimizing link transmission delay and link load, optimize and determine the resource allocation strategy of user flow at L _{i, j}

Further, in step S9, the network slice and resource allocation strategy of the user flow k is

The beneficial effect of the present invention is that: the present invention can effectively reduce network end-to-end delay, balance network load, and improve network stability.

Description of drawings

In order to make the purpose, technical scheme and beneficial effect of the present invention clearer, the present invention provides the following drawings for illustration:

FIG. 1 is a block diagram of an exemplary communication system according to an embodiment of the present invention;

FIG. 2 is a flowchart of joint routing selection and resource allocation according to an embodiment of the present invention;

FIG. 3 is a communication schematic diagram of an application joint routing selection and resource allocation method according to an embodiment of the present invention.

Detailed ways

The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

SDN uses OpenFlow to separate the control function from the forwarding function, and the controller completes the network control function. OpenFlow switches store flow tables to realize data flow lookup and forwarding. The switch is connected to the external controller through the secure channel through the OpenFlow protocol to query and manage the flow table. The flow table includes header fields (Header fields, matching multiple header fields), activity counters (Counters), and 0 or more execution actions (Actions). The OpenFlow switch searches the data flow, and if it matches, executes the relevant policy; otherwise, forwards the data flow to the controller through a secure channel, and the controller executes the corresponding decision.

FIG. 1 is a block diagram of an exemplary communication system according to an embodiment of the present invention. In the communication system, the SDN control plane includes a centralized controller; the SDN data plane includes multiple distributed forwarding devices, where the forwarding devices can be any routers, switches, virtual switches, etc. that support the SDN protocol.

FIG. 2 is a flowchart of joint routing selection and resource allocation according to an embodiment of the present invention. The method in the embodiment comprises the following steps:

S201. The controller receives a user stream transmission request;

S202. Determine a candidate switch set and a candidate link set;

S203, modeling network slicing and resource allocation identification;

S204. Determine network slicing and resource allocation limiting conditions;

S205, modeling user flow end-to-end delay and network load;

S206, aiming at minimizing the user flow end-to-end delay and network load, and determining a network slicing strategy for each user flow optimization;

S207, determine whether there are shared switches and shared links in each network slice, if there are shared switches and shared links, execute S208; otherwise, execute S209;

S208, judging whether the shared switch and the shared link meet the resource constraints of the switch and the link, and if so, execute S209; otherwise, execute a resource allocation strategy for the shared switch and the shared link;

S209, the controller delivers the network slicing and resource allocation policies to corresponding switches, and each switch completes user flow routing and resource allocation according to the flow table.

FIG. 3 is a schematic diagram of communication of an application joint routing selection and resource allocation method according to an embodiment of the present invention. The routing selection process is described in detail below in conjunction with the embodiment:

(1) The data flows f ₁ and f ₂ send routing requests to the source switches S ₁ and S ₂ respectively.

(2) The source switches S ₁ and S ₂ find that there is no matching flow entry in the local flow table, and execute S201 to send a routing request message to the controller.

(3) The controller executes S202 to determine the set of candidate switches for _f1 set with candidate links Determine the set of candidate switches for _f2 set with candidate links

(4) The controller executes S203, modeling network slices and resource allocation identifiers.

(5) The controller executes S204 to determine network slices and resource allocation constraints.

(6) The controller executes S205 to model the user flow end-to-end delay and network load, then the objective function can be obtained and

(7) The controller executes step S206 to determine the _f1 network slicing strategy as Set the resource allocation policy of f ₁ to f ₂ The network slicing strategy is Set the resource allocation policy of f ₂ to

(8) The controller executes step S207, and judges that there is a shared node N ₃ between f ₁ and f ₂ and there is no shared link, and then executes S208.

(9) The controller executes step S208 to judge whether the shared node _N3 satisfies the node resource constraint, and if the shared node _N3 satisfies the node resource constraint, then executes S210.

(10) Each switch executes S210 to complete the routing and resource allocation of _f1 and _f2 .

Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that it can be described in terms of form and Various changes may be made in the details without departing from the scope of the invention defined by the claims.

Claims (10)

1. An SDN joint routing and resource allocation method based on network slices is characterized in that: the method comprises the following steps:

s1: a Software Defined Network (SDN) controller receiving a user streaming request;

s2: the controller determines a candidate switch set and a candidate link set according to the resource requirements of the user flow;

s3: modeling network slices and resource allocation identifications;

s4: determining a network slice and a resource allocation limiting condition;

s5: modeling end-to-end time delay of user flow and network load;

s6: optimizing and determining a network slicing strategy for each user flow by taking the end-to-end time delay of the user flow and the minimization of network load as targets;

s7: judging whether each network slice has a shared switch and a shared link, if so, executing the step S8; otherwise, go to step S9;

s8: judging whether the shared switch and the shared link meet the switch and link resource constraint, if so, executing the step S9; otherwise, executing resource allocation strategy for the shared switch and the shared link;

s9: the controller issues the network slice and the resource allocation strategy to the corresponding switches, and each switch completes the routing and resource allocation of the user flow according to the flow table.

2. The SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S1, the user streaming request is modeled asWherein S is_kAnd D_kSource and destination switches respectively representing kth subscriber flow, f_kIndicating the data traffic of the k-th user flow,indicating the lowest transmission rate requirement for the k-th user stream,andthe switch computation resource demand and the storage resource demand of the kth user flow are respectively represented.

3. According to claim1, the SDN joint routing and resource allocation method based on the network slice is characterized in that: in step S2, the candidate switch set is determined according to the switch resource requirement of the user flow, i.e. N_iAs a candidate switch for user flow k, only if the computing resources of that switchAnd storage resourcesSatisfy the resource requirements of user flow k, i.e.And isTo be provided withSet of candidate switches representing the kth subscriber flow, i.e.

The candidate link set is determined according to the link resource requirements of the user flow, i.e. L_i,jAs a candidate link for user flow k, only if the bandwidth resource B of the link is available_i,jThe bandwidth resource requirement of the user flow k is satisfied, and the sending rate R of the link is_i,jMeeting the minimum sending rate requirement of user flow kI.e. f_k≤B_i,jAnd isTo be provided withTo representCandidate link sets for the k-th user flow, i.e.

4. The SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S3, the modeling network slice and the resource allocation identifier are specifically: let x_i,j,kE {0,1} is the network slice ID of user flow k, if x_i,j,kIf 1, it means that the network slice allocated for the user flow k includes the switch N_iAnd link L_i,jOtherwise, x_i,j,k＝0；Respectively representing the occupation of N by user flow k_iCalculating the resource proportion and the storage resource proportion; delta_i,j,k∈[0,1]Indicating that user stream k occupies L_i,jThe transmission resource ratio of (1).

5. The SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S4, the network slice and the resource allocation limiting condition include switch flow conservation, and a switch, link capacity and a resource allocation limiting condition, where the switch flow conservation condition is specifically as follows:

routing source switch S for user flow k_kOnly one output path is selected, i.e.

Trunk exchanger N for user flow k_jAt equal input and output flows, i.e.

Destination switch D for user flow k_kOnly one of them is selectedInput paths, i.e.

The switch capacity limiting condition comprises a switch computing resource capacity and storage resource capacity limiting condition, and the modeling is as follows:

user flow occupation N_iCannot exceed the amount of computing resources of the switch, i.e. the switch is not powered onUser flow occupation N_iCannot exceed the amount of storage resources of the switch, i.e.

The limiting condition of the link capacity is that the user flow occupies L_i,jCannot exceed the amount of bandwidth resources of the link, i.e. the link resources of

The switch resource allocation limiting condition comprises a switch computing resource and storage resource allocation limiting condition and is modeled as follows:

user flow occupation N_iIs not more than 1, i.e. the sum of the allocation proportions of the computing resources of

User flow occupation N_iIs not more than 1, i.e. the sum of the allocation proportions of the storage resources

The limited condition of link resource allocation is that the user flow occupies L_i,jIs not more than 1, i.e. the sum of the bandwidth resource allocation proportions

6. The SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S5, the system end-to-end delay is modeled as the sum of the link transmission delay and the switch queuing delay, i.e. the system end-to-end delay is modeled as the sum of the link transmission delay and the switch queuing delayWherein D is_i,jIndicating the link transmission delay, D_iIndicating queuing delay of switch, for D_i,jAnd D_iThe modeling is as follows:

modeling link transmission delay

The modeling exchanger processes the data process into an M | M |1 queuing system to obtainWherein, mu_iRepresents N_iService rate of (λ)_iRepresents N_iThe service arrival rate of (2);

the network load is modeled as the sum of the link load and the switch load, namely:wherein, U_i,jIndicating the link load, U_iIndicating switch load, for U_i,jAnd U_iThe modeling is as follows:

modeling link load

Modeling switch loadWherein,the load is calculated for the switch and,store the load for the switch, α are weight coefficients.

7. The SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S6, the network slicing policy is to determine a network slicing policy for each user flow respectively under the condition that the network slicing and resource allocation limiting conditions are satisfied, that is, for each user flow k, orderThe resource allocation strategy for each flow is optimized and determined by taking the end-to-end time delay of the user flow and the network load minimization as the targets, namely:

8. the SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S7, the determination of whether or not there are shared switches and shared links in each network slice is to determineAnd if so, a shared switch and a shared link exist.

9. The SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S8, the determining whether the shared switch and the shared link satisfy the switch and link resource constraints includes:

judging switch N_iIf the user flow occupies more than the computing resource of the switch, the computing resource of the switch is constrained, that is, the user flow occupies more than the computing resource of the switchOptimizing and determining the user flow in N by using the minimization of the queuing delay of the switch and the calculation load of the switch as the target_jComputing resource allocation policy of

Judging switch N_iIf the storage resources of the switch are occupied by multiple user flows and exceed the storage resource amount, that isOptimizing and determining the user flow at N by using the minimization of the storage load of the switch as a target_jStorage resource allocation policy of

Judging the link L_i,jIf the bandwidth resource occupied by the multiple user streams exceeds the link bandwidth resource amount, that isOptimally determining the user flow at L by using the link transmission delay and the link load minimization as the target_i,jResource allocation policy of

10. The SDN federated routing and resource allocation method based on network slice of claim 1, wherein: in step S9, the network slice and the resource allocation policy of the user flow k are